Networking Domain Name System

Transcription

1 System i Networking Domain Name System Version 6 Release 1

2

3 System i Networking Domain Name System Version 6 Release 1

4 Note Before using this information and the product it supports, read the information in Notices, on page 43. This edition applies to version 6, release 1, modification 0 of IBM i5/os (product number 5761-SS1) and to all subsequent releases and modifications until otherwise indicated in new editions. This version does not run on all reduced instruction set computer (RISC) models nor does it run on CISC models. Copyright IBM Corporation 1998, US Government Users Restricted Rights Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.

5 Contents Domain Name System PDF file for Domain Name System Domain Name System concepts Understanding zones Understanding Domain Name System queries.. 3 Domain Name System domain setup Dynamic updates BIND 9 features Domain Name System resource records Mail and Mail Exchanger records Examples: Domain Name System Example: Single Domain Name System server for an intranet Example: Single Domain Name System server with Internet access Example: Domain Name System and Dynamic Host Configuration Protocol on the same System i Example: Splitting DNS over firewall by setting up two DNS servers on the same System i Example: Splitting DNS over firewall by using view Planning for Domain Name System Determining Domain Name System authorities 24 Determining domain structure Planning security measures Domain Name System requirements Determining if Domain Name System is installed 26 Installing Domain Name System Configuring Domain Name System Accessing Domain Name System in System i Navigator Configuring name servers Creating a name server instance Editing Domain Name System server properties Configuring zones on a name server Configuring views on a name server Configuring Domain Name System to receive dynamic updates Importing Domain Name System files Record validation Accessing external Domain Name System data 30 Managing Domain Name System Verifying the Domain Name System function is working Managing security keys Managing Domain Name System keys Managing dynamic update keys Accessing Domain Name System server statistics 33 Accessing server statistics Accessing an active server database Maintaining Domain Name System configuration files Advanced Domain Name System features Starting or stopping Domain Name System servers Changing debug values Troubleshooting Domain Name System Logging Domain Name System server messages 37 Changing Domain Name System debug settings 40 Related information for Domain Name System.. 40 Appendix. Notices Programming interface information Trademarks Terms and conditions Copyright IBM Corp. 1998, 2008 iii

6 iv System i: Networking Domain Name System

7 Domain Name System Domain Name System (DNS) is a distributed database system for managing host names and their associated Internet Protocol (IP) addresses. With DNS, people can use simple names, such as to locate a host, rather than using the IP addresses, for example, in IPv4, or 2001:D88::1 in IPv6. A single server might be responsible only for knowing the host names and IP addresses for a small part of a zone, but DNS servers can work together to map all domain names to their IP addresses. DNS servers that work together allow computers to communicate across the Internet. For IBM i Version 6 Release 1 (V6R1), DNS services are based on the industry-standard DNS implementation, known as Berkeley Internet Name Domain (BIND) version 9. For previous IBM i releases, DNS services were based on BIND version To use the new BIND version 9 DNS server, you must have IBM i option 31 (DNS) and option 33 (Portable Application Solutions Environment (PASE)) installed. As of IBM i V6R1, for security reasons, BIND 4 and 8 are replaced with BIND 9. Thus, the migration to BIND 9 is required for your DNS server. PDF file for Domain Name System You can view and print a PDF file of this information. To view or download the PDF version of this document, select Domain Name System (about 625 KB). Saving PDF files To save a PDF on your workstation for viewing or printing: 1. Right-click the PDF link in your browser. 2. Click the option that saves the PDF locally. 3. Navigate to the directory in which you want to save the PDF. 4. Click Save. Downloading Adobe Reader You need Adobe Reader installed on your system to view or print these PDFs. You can download a free copy from the Adobe Web site ( Related reference: Related information for Domain Name System on page 40 IBM Redbooks publications, Web sites, and other information center topic collections contain information that relates to the Domain Name System (DNS) topic collection. You can view or print any of the PDF files. Domain Name System concepts Domain Name System (DNS) is a distributed database system for managing host names and their associated Internet Protocol (IP) addresses. With DNS, you can use simple names, such as to locate a host, rather than using the IP addresses, for example, in IPv4, or 2001:D88::1 in IPv6. Copyright IBM Corp. 1998,

8 A single server might be responsible only for knowing the host names and IP addresses for a small part of a zone, but DNS servers can work together to map all domain names to their IP addresses. DNS servers that work together allows computers to communicate across the Internet. DNS data is broken up into a hierarchy of domains. Servers are responsible to know only a small portion of data, such as a single subdomain. The portion of a domain for which the server is directly responsible is called a zone. A DNS server that has complete host information and data for a zone is authoritative for the zone. An authoritative server can answer queries about hosts in its zone, using its own resource records. The query process depends on a number of factors. Understanding DNS queries explains the paths that a client can use to resolve a query. Understanding zones Domain Name System (DNS) data is divided into manageable sets of data called zones. And each of these sets is a specific zone type. Zones contain name and IP address information about one or more parts of a DNS domain. A server that contains all of the information for a zone is the authoritative server for the domain, called a parent zone. Sometimes it makes sense to delegate the authority for answering DNS queries for a particular subdomain to another DNS server, called a child zone. In this case, the DNS server for the domain can be configured to refer the subdomain queries to the appropriate server. For backup and redundancy, zone data is often stored on servers other than the authoritative DNS server. These other servers are called secondary servers, which load zone data from the authoritative server. Configuring secondary servers allows you to balance the demand on servers and also provides a backup in case the primary server goes down. Secondary servers obtain zone data by doing zone transfers from the authoritative server. When a secondary server is initialized, it loads a complete copy of the zone data from the primary server. The secondary server also reloads zone data from the primary server or from other secondaries for that domain when zone data changes. DNS zone types You can use IBM i DNS to define several types of zones to help you manage DNS data: Primary zone A primary zone loads zone data directly from a file on a host. It can contain a subzone, or child zone. It can also contain resource records, such as host, alias (CNAME), IPv4 address (A), IPv6 address (AAAA), or reverse mapping pointer (PTR) records. Note: Primary zones are sometimes referred to as master zones in other BIND documentation. Subzone A subzone defines a zone within the primary zone. Subzones allow you to organize zone data into manageable pieces. Child zone A child zone defines a subzone and delegates responsibility for the subzone data to one or more name servers. Alias (CNAME) An alias defines an alternate name for a primary domain name. Host A host object maps A and PTR records to a host. Additional resource records can be associated with a host. Secondary zone A secondary zone loads zone data from a zone's primary server or another secondary server. It maintains a complete copy of the zone for which it is a secondary. 2 System i: Networking Domain Name System

9 Note: Secondary zones are sometimes referred to as slave zones in other BIND documentation. Stub zone A stub zone is similar to a secondary zone, but it only transfers the name server (NS) records for that zone. Forward zone A forward zone directs all queries for that particular zone to other servers. Related concepts: Understanding Domain Name System queries Domain Name System (DNS) clients use DNS servers to resolve queries. The queries might come directly from the client or from an application running on the client. Related tasks: Configuring zones on a name server on page 28 After you configure a Domain Name System (DNS) server instance, you need to configure the zones for the name server. Related reference: Example: Single Domain Name System server for an intranet on page 14 This example depicts a simple subnet with a Domain Name System (DNS) server for internal use. Domain Name System resource records on page 9 Resource records are used to store data about domain names and IP addresses. You can use the Resource record lookup table to look into the resource records supported for the IBM i operating system. Understanding Domain Name System queries Domain Name System (DNS) clients use DNS servers to resolve queries. The queries might come directly from the client or from an application running on the client. The client sends a query message to the DNS server that contains a fully qualified domain name (FQDN), a query type, such as a particular resource record the client requires, and the class for the domain name, which is typically the Internet (IN) class. The following figure depicts the sample network from the Single DNS server with Internet access example. Domain Name System 3

10 Figure 1. Single DNS server with Internet access Suppose that host dataentry queries the DNS server for graphics.mycompany.com. The DNS server uses its own zone data and responds with the IP address Now suppose dataentry requests the IP address of This host is not in the DNS server's zone data. Two paths can be followed: recursion or iteration. If a DNS server is set to use recursion, the server can query or contact other DNS servers on behalf of the requesting client to fully resolve the name, and then send an answer back to the client. Additionally, the requesting server stores the answer into its cache so that the answer can be used the next time that the server receives that query. If a DNS server is set to use iteration, a client can attempt to contact other DNS servers on its own to resolve a name. In this process, the client uses separate and additional queries based on referral answers from servers. 4 System i: Networking Domain Name System

11 Related reference: Understanding zones on page 2 Domain Name System (DNS) data is divided into manageable sets of data called zones. And each of these sets is a specific zone type. Example: Single Domain Name System server with Internet access on page 16 This example depicts a simple subnet with a Domain Name System (DNS) server connected directly to the Internet. Domain Name System domain setup Domain Name System (DNS) domain setup requires domain name registration to prevent others from using your domain name. DNS allows you to serve names and addresses on an intranet, or internal network. It also allows you to serve names and addresses to the rest of the world through the Internet. If you want to set up domains on the Internet, you are required to register a domain name. If you are setting up an intranet, you are not required to register a domain name for internal use. Whether to register an intranet name depends on whether you want to ensure that no one else can ever use the name on the Internet, independent of your internal use. Registering a name that you are going to use internally ensures that you will never have a conflict if you later want to use the domain name externally. Domain registration can be performed by direct contact with an authorized domain name registrar, or through some Internet Service Providers (ISPs). Some ISPs offer a service to submit domain name registration requests on your behalf. The Internet Network Information Center (InterNIC) maintains a directory of all domain name registrars that are authorized by the Internet Corporation for Assigned Names and Numbers (ICANN). Related reference: Example: Single Domain Name System server with Internet access on page 16 This example depicts a simple subnet with a Domain Name System (DNS) server connected directly to the Internet. Related information: Internet Network Information Center (InterNIC) Dynamic updates IBM i Domain Name System (DNS) that is based on BIND 9 supports dynamic updates. Outside sources, such as Dynamic Host Configuration Protocol (DHCP), can send updates to the DNS server. In addition, you can also use DNS client tools, such as Dynamic Update Utility (NSUPDATE), to perform dynamic updates. DHCP is a TCP/IP standard that uses a central server to manage IP addresses and other configuration details for an entire network. A DHCP server responds to requests from clients, dynamically assigning properties to them. DHCP allows you to define network host configuration parameters at a central location and automate the configuration of hosts. It is often used to assign temporary IP addresses to clients for networks that contain more clients than the number of IP addresses available. In the past, all DNS data was stored in static databases. All DNS resource records had to be created and maintained by the administrator. But, DNS servers that are based on BIND 8, or later, can be configured to accept requests from other sources to update zone data dynamically. You can configure your DHCP server to send update requests to the DNS server each time it assigns a new address to a host. This automated process reduces DNS server administration in rapidly growing or changing TCP/IP networks, and in networks where hosts change locations frequently. When a client Domain Name System 5

12 using DHCP receives an IP address, that data is immediately sent to the DNS server. Using this method, DNS can continue to successfully resolve queries for hosts, even when their IP addresses change. You can configure DHCP to update address mapping (A for IPv4 or AAAA for IPv6) records, reverse-lookup pointer (PTR) records, or both on behalf of a client. The address mapping record (A or AAAA) maps a machine's host name to its IP address. The PTR record maps a machine's IP address to its host name. When a client's address changes, DHCP can automatically send an update to the DNS server so other hosts in the network can locate the client through DNS queries at the client's new IP address. For each record that is updated dynamically, an associated Text (TXT) record is written to identify that the record was written by DHCP. Note: If you set DHCP to update only PTR records, you must configure DNS to allow updates from clients so that every client can update its A record if the client uses IPv4 address, or update its AAAA record if the client uses IPv6 address. Not all DHCP clients support making their own A or AAAA record update requests. Consult the documentation for your client platform before choosing this method. Dynamic zones are secured by creating a list of authorized sources that are allowed to send updates. You can define authorized sources using individual IP addresses, whole subnets, packets that have been signed using a shared secret key (called a Transaction Signature, or TSIG), or any combination of those methods. DNS verifies that incoming request packets are coming from an authorized source before updating the resource records. Dynamic updates can be performed between DNS and DHCP on a single System i platform, between different System i platforms, or between a System i platform and other systems that are capable of dynamic updates. Note: The dynamic Update DNS (QTOBUPDT) API is required on servers that are sending dynamic updates to DNS. It is installed automatically with IBM i Option 31, DNS. However, in BIND 9, the NSUPDATE command is the preferred method to make the updates on the System i platform. 6 System i: Networking Domain Name System

13 Related concepts: Dynamic Host Configuration Protocol Related tasks: Configuring Domain Name System to receive dynamic updates on page 29 Domain Name System (DNS) servers running BIND 9 can be configured to accept requests from other sources to update zone data dynamically. This topic provides instructions for configuring the allow-update option so DNS can receive dynamic updates. Configuring the DHCP to send dynamic updates to DNS Related reference: Example: Domain Name System and Dynamic Host Configuration Protocol on the same System i on page 18 This example depicts Domain Name System (DNS) and Dynamic Host Configuration Protocol (DHCP) on the same System i platform. Domain Name System resource records on page 9 Resource records are used to store data about domain names and IP addresses. You can use the Resource record lookup table to look into the resource records supported for the IBM i operating system. QTOBUPT BIND 9 features BIND 9 is similar to BIND 8; however, it provides several features to enhance performance of your Domain Name System (DNS) server, such as views. BIND 9 features BIND 9 is similar to BIND 8; however, it provides several features to enhance performance of your Domain Name System (DNS) server, such as views. Views on a single IBM i DNS server The view statement allows a single DNS instance to answer a query differently depending on where the query comes from, such as the Internet or an intranet. One practical application of the view feature is to split DNS setups without having to run multiple DNS servers. For example, in a single DNS server, you can define a view to answer queries from an internal network, while define another view to answer queries from external network. New client commands The following client commands enhance the management capability of your DNS server: Dynamic Update Utility (NSUPDATE) The Dynamic Update Utility (NSUPDATE) command is used to submit Dynamic DNS Update requests as defined in Request for Comments (RFC) 2136 to a DNS server. This allows resource records to be added or removed from a zone while the DNS server is running. Thus, you do not need to update records by manually editing the zone file. A single update request can contain requests to add or remove multiple resource records, but the resource records that are dynamically added or removed with the NSUPDATE command should be in the same zone. Note: Do not manually edit zones that are under dynamic control by using the NSUPDATE command or through a DHCP server. Manual edits might conflict with dynamic updates and cause data to be lost. Start DIG Query (DIG) Domain Information Groper (DIG) is a more powerful query tool, compared with the Name Server Lookup (NSLOOKUP) command, that you can use to retrieve information from a DNS server or test the response of a DNS server. The NSLOOKUP command is deprecated and is only provided Domain Name System 7

14 for compatibility with earlier versions. You can use DIG to verify that a DNS server is responding correctly before you configure your system to use it. You can also retrieve DNS information about hosts, domains, and other DNS servers by using DIG. You can use the Start DIG Query (STRDIGQRY) command or its alias DIG to start the Domain Information Groper tool. Start HOST Query (HOST) The Start HOST Query (HOST) command is used for DNS lookups. You can use it to convert domain names to IP addresses (either IPv4 or IPv6) and vice versa. Remote Name Daemon Control (RNDC) The Remote Name Daemon Control (RNDC) command is a powerful utility that allows a system administrator to control the operation of a name server. It reads a configuration file, called rndc.conf, to determine how to contact the name server and to determine what algorithm and key it should use. If no rndc.conf file is found, then, by default, an rndc-key._kid file that is created during installation is used, which automatically grants access through the loopback interface. IPv6 support BIND 9 supports name-to-address and address-to-name lookups in all currently defined forms of IPv6. For forward lookups, BIND 9 supports both AAAA and A6 records, but A6 records are now deprecated. For IPv6 reverse lookups, it supports the traditional nibble format used in the ip6.arpa domain, as well as the older, deprecated ip6.int domain. Journal files Journal files are used to hold dynamic updates for a zone. It is automatically created when the first dynamic update from a client is received, and does not disappear. This is a binary file and should not be edited. With the journal file, when a server is restarted after a shutdown or crash, it replays the journal file to incorporate into the zone any updates that took place after the last zone dump. Journal files are also used to store updates for the incremental zone transfers (IXFR) method. DNS for IBM i has been redesigned to use BIND 9. To run BIND 9 DNS on your system, your system must meet certain software requirements. 8 System i: Networking Domain Name System

15 Related concepts: Domain Name System requirements on page 26 Consider these software requirements to run Domain Name System (DNS) on your System i platform. Dynamic updates on page 5 IBM i Domain Name System (DNS) that is based on BIND 9 supports dynamic updates. Outside sources, such as Dynamic Host Configuration Protocol (DHCP), can send updates to the DNS server. In addition, you can also use DNS client tools, such as Dynamic Update Utility (NSUPDATE), to perform dynamic updates. Related reference: Example: Splitting DNS over firewall by setting up two DNS servers on the same System i on page 20 This example depicts a Domain Name System (DNS) server that operates over a firewall to protect internal data from the Internet, while allowing internal users to access data on the Internet. This configuration accomplishes this protection by setting up two DNS servers on the same System i platform. Planning security measures on page 25 Domain Name System (DNS) provides security options to limit outside access to your server. Domain Name System resource records Resource records are used to store data about domain names and IP addresses. You can use the Resource record lookup table to look into the resource records supported for the IBM i operating system. A DNS zone database is made up of a collection of resource records. Each resource record specifies information about a particular object. For example, address mapping (A) records map a host name to an IP address, and reverse-lookup pointer (PTR) records map an IP address to a host name. The server uses these records to answer queries for hosts in its zone. For more information, use the table to view DNS resource records. Note: The entries in the resource record lookup table might be added or removed according to the change of the BIND document. Also, this is not a comprehensive list of all resource records listed in BIND. Table 1. Resource record lookup table Resource record Abbreviation Description Address Mapping records A The A record specifies the IP address of this host. A records are used to resolve a query for the IP address of a specific domain name. This record type is defined in Request for Comments (RFC) Andrew File System Database records AFSDB The AFSDB record specifies the AFS or DCE address of the object. AFSDB records are used like A records to map a domain name to its AFSDB address; or to map from the domain name of a cell to authenticated name servers for that cell. This record type is defined in RFC Canonical Name records CNAME The CNAME record specifies the actual domain name of this object. When DNS queries an aliased name and finds a CNAME record pointing to the canonical name, it then queries that canonical domain name. This record type is defined in RFC Domain Name System 9

16 Table 1. Resource record lookup table (continued) Resource record Abbreviation Description Host Information records HINFO The HINFO record specifies general information about a host. Standard CPU and operating system names are defined in the Assigned Numbers RFC However, use of the standard numbers is not required. This record type is defined in RFC Integrated Services Digital Network records ISDN The ISDN record specifies the address of this object. This record maps a host name to the ISDN address. They are used only in ISDN networks. This record type is defined in RFC IP Version 6 Address records AAAA The AAAA record specifies the 128-bit IPv6 address of a host. AAAA records, which are similar to A records, are used to resolve queries for the IPv6 address of a specific domain name. This record type is defined in RFC Location records LOC The LOC record specifies the physical location of network components. These records can be used by applications to evaluate network efficiency or map the physical network. This record type is defined in RFC Mail Exchanger records MX The MX records defines a mail exchanger host for mail sent to this domain. These records are used by Simple Mail Transfer Protocol (SMTP) to locate hosts that processes or forwards mail for this domain, along with preference values for each mail exchanger host. Each mail exchanger host must have a corresponding host address (A) records in a valid zone. This record type is defined in RFC Mail Group records MG The MG records specifies the mail group domain name. This record type is defined in RFC Mailbox records MB The MB records specifies the host domain name which contains the mailbox for this object. Mail sent to the domain is directed to the host specified in the MB record. This record type is defined in RFC System i: Networking Domain Name System

17 Table 1. Resource record lookup table (continued) Resource record Abbreviation Description Mailbox Information records MINFO The MINFO records specifies the mailbox that should receive messages or errors for this object. The MINFO record is more commonly used for mailing lists than for a single mailbox. This record type is defined in RFC Mailbox Rename records MR The MR records specifies a new domain name for a mailbox. Use the MR record as a forwarding entry for a user who has moved to a different mailbox. This record type is defined in RFC Name Server records NS The NS record specifies an authoritative name server for this host. This record type is defined in RFC Network Service Access Protocol records NSAP The NSAP record specifies the address of a NSAP resource. NSAP records are used to map domain names to NSAP addresses. This record type is defined in RFC Public Key records KEY The KEY record specifies a public key that is associated with a DNS name. The key can be for a zone, a user, or a host. This record type is defined in RFC Responsible Person records RP The RP record specifies the internet mail address and description of the person responsible for this zone or host. This record type is defined in RFC Reverse-lookup Pointer records PTR The PTR record specifies the domain name of a host for which you want a PTR record defined. PTR records allow a host name lookup, given an IP address. This record type is defined in RFC Route Through records RT The RT record specifies a host domain name that can act as a forwarder of IP packets for this host. This record type is defined in RFC Services records SRV The SRV record specifies the hosts that support the defined services in the record. This record type is defined in RFC Domain Name System 11

18 Table 1. Resource record lookup table (continued) Resource record Abbreviation Description Start of Authority records SOA The SOA record specifies that this server is authoritative for this zone. An authoritative server is the best source for data within a zone. The SOA record contains general information about the zone and reload rules for secondary servers. There can be only one SOA record per zone. This record type is defined in RFC Text records TXT The TXT record specifies multiple strings of text, up to 255 characters long each, to be associated with a domain name. TXT records can be used along with responsible person (RP) records to provide information about who is responsible for a zone. This record type is defined in RFC TXT records are used by IBM i DHCP for dynamic updates. The DHCP server writes an associated TXT record for each PTR and an A record update that is done by the DHCP server. DHCP records have a prefix of AS400DHCP. Well-Known Services records WKS The WKS record specifies the well-known services supported by the object. Most commonly, WKS records indicate whether tcp or udp or both protocols are supported for this address. This record type is defined in RFC X.400 Address Mapping records PX The PX records is a pointer to X.400/RFC 822 mapping information. This record type is defined in RFC X25 Address Mapping records X25 The X25 record specifies the address of an X25 resource. This record maps a host name to the PSDN address. They are used only in X25 networks. This record type is defined in RFC System i: Networking Domain Name System

19 Related concepts: Mail and Mail Exchanger records Domain Name System (DNS) supports advanced mail routing through the use of Mail and Mail Exchanger (MX) records. Related reference: Example: Single Domain Name System server for an intranet on page 14 This example depicts a simple subnet with a Domain Name System (DNS) server for internal use. Understanding zones on page 2 Domain Name System (DNS) data is divided into manageable sets of data called zones. And each of these sets is a specific zone type. Mail and Mail Exchanger records Domain Name System (DNS) supports advanced mail routing through the use of Mail and Mail Exchanger (MX) records. Mail and MX records are used by mail routing programs, such as Simple Mail Transfer Protocol (SMTP). The lookup table in DNS resource records contains the types of mail records that IBM i DNS supports. DNS includes information for sending electronic mail by using mail exchanger information. If the network is using DNS, an SMTP application does not deliver mail addressed to host TEST.IBM.COM by opening a TCP connection to TEST.IBM.COM. SMTP first queries the DNS server to find out which host servers can be used to deliver the message. Deliver mail to a specific address DNS servers use resource records that are known as mail exchanger (MX) records. MX records map a domain or host name to a preference value and host name. MX records are generally used to designate that one host is used to process mail for another host. The records are also used to designate another host to deliver mail to, if the first host cannot be reached. In other words, they allow a mail that is addressed to one host to be delivered to a different host. Multiple MX resource records might exist for the same domain or host name. When multiple MX records exist for the same domain or host, the preference (or priority) value of each record determines the order in which they are tried. The lowest preference value corresponds to the most preferred record, which is tried first. When the most preferred host cannot be reached, the sending mail application tries to contact the next, less preferred MX host. The domain administrator, or the creator of the MX record, sets the preference value. A DNS server can respond with an empty list of MX resource records when the name is in the DNS server's authority but has no MX assigned to it. When this occurs, the sending mail application might try to establish a connection with the destination host directly. Note: Using a wildcard (example: *.mycompany.com) in MX records for a domain is not suggested. Example: MX record for a host In the following example, the system, by preference, delivers mail for fsc5.test.ibm.com to the host itself. If the host cannot be reached, the system might deliver the mail to psfred.test.ibm.com or to mvs.test.ibm.com (if psfred.test.ibm.com also cannot be reached). This is an example of what these MX records will look like: fsc5.test.ibm.com IN MX 0 fsc5.test.ibm.com IN MX 2 psfred.test.ibm.com IN MX 4 mvs.test.ibm.com Domain Name System 13

20 Related reference: Domain Name System resource records on page 9 Resource records are used to store data about domain names and IP addresses. You can use the Resource record lookup table to look into the resource records supported for the IBM i operating system. Examples: Domain Name System You can use these examples to understand how to use Domain Name System (DNS) in your network. DNS is a distributed database system for managing host names and their associated IP addresses. The following examples help to explain how DNS works, and how you can use it in your network. The examples describe the setup and reasons it will be used. They also link to related concepts that you might find useful to understand the pictures. Example: Single Domain Name System server for an intranet This example depicts a simple subnet with a Domain Name System (DNS) server for internal use. The following figure depicts DNS running on a System i platform for an internal network. This single DNS server instance is set up to listen for queries on all interface IP addresses. The system is a primary name server for the mycompany.com zone. 14 System i: Networking Domain Name System

21 Figure 2. Single DNS server for an intranet Each host in the zone has an IP address and a domain name. The administrator must manually define the hosts in the DNS zone data by creating resource records. Address mapping records (A for IPv4 or AAAA for IPv6) map the name of a machine to its associated IP address. This allows other hosts on the network to query the DNS server to find the IP address assigned to a particular host name. Reverse-lookup pointer (PTR) records map the IP address of a machine to its associated name. This allows other hosts on the network to query the DNS server to find the host name that corresponds to an IP address. In addition to A, AAAA, and PTR records, DNS supports many other resource records that might be required, depending on what other TCP/IP-based applications you are running on your intranet. For example, if you are running internal systems, you might need to add mail exchanger (MX) records so that SMTP can query DNS to find out which systems are running the mail servers. If this small network were part of a larger intranet, it might be necessary to define internal root servers. Domain Name System 15

22 Secondary servers Secondary servers load zone data from the authoritative server. Secondary servers obtain zone data by doing zone transfers from the authoritative server. When a secondary name server starts, it requests all data for the specified domain from the primary name server. A secondary name server requests updated data from the primary server either because it receives notification from the primary name server (if the NOTIFY function is being used) or because it queries the primary name server and determines that the data has changed. In the figure above, the mysystemi server is part of an intranet. Another system, mysystemi2, has been configured to act as a secondary DNS server for the mycompany.com zone. The secondary server can be used to balance the demand on servers and also to provide a backup in case the primary server goes down. It is a good practice to have at least one secondary server for every zone. Related reference: Domain Name System resource records on page 9 Resource records are used to store data about domain names and IP addresses. You can use the Resource record lookup table to look into the resource records supported for the IBM i operating system. Understanding zones on page 2 Domain Name System (DNS) data is divided into manageable sets of data called zones. And each of these sets is a specific zone type. Example: Single Domain Name System server with Internet access This example depicts a simple subnet with a Domain Name System (DNS) server connected directly to the Internet. Example: Single Domain Name System server with Internet access This example depicts a simple subnet with a Domain Name System (DNS) server connected directly to the Internet. The following figure depicts the same example network from the single DNS server for intranet example, but now the company has added a connection to the Internet. In this example, the company is able to access the Internet, but the firewall is configured to block Internet traffic into the network. 16 System i: Networking Domain Name System

23 Figure 3. Single DNS server with Internet access To resolve Internet addresses, you need to do at least one of the following tasks: v Define Internet root servers You can load the default Internet root servers automatically, but you might need to update the list. These servers can help to resolve addresses outside of your own zone. For instructions for obtaining the current Internet root servers, see Accessing external Domain Name System data. v Enable forwarding You can set up forwarding to pass queries for zones outside of mycompany.com to external DNS servers, such as DNS servers run by your Internet service provider (ISP). If you want to enable Domain Name System 17

24 searching by both forwarding and root servers, you need to set the forward option to first. The server first tries forwarding and then queries the root servers only if forwarding fails to resolve the query. The following configuration changes might also be required: v Assign unrestricted IP addresses In the example above, 10.x.x.x addresses are shown. However, these are restricted addresses and cannot be used outside of an intranet. They are shown below for example purposes, but your own IP addresses is determined by your ISP and other networking factors. v Register your domain name If you are visible to the Internet and have not already registered, you need to register a domain name. v Establish a firewall It is not suggested that you allow your DNS to be directly connected to the Internet. You need to configure a firewall or take other precautions to secure your System i platform. Related concepts: Domain Name System domain setup on page 5 Domain Name System (DNS) domain setup requires domain name registration to prevent others from using your domain name. System i and Internet security Understanding Domain Name System queries on page 3 Domain Name System (DNS) clients use DNS servers to resolve queries. The queries might come directly from the client or from an application running on the client. Related reference: Example: Single Domain Name System server for an intranet on page 14 This example depicts a simple subnet with a Domain Name System (DNS) server for internal use. Example: Domain Name System and Dynamic Host Configuration Protocol on the same System i This example depicts Domain Name System (DNS) and Dynamic Host Configuration Protocol (DHCP) on the same System i platform. The configuration can be used to update DNS zone data dynamically when DHCP assigns IP addresses to hosts. The following figure depicts a small subnet network with one System i platform that acts as a DHCP and DNS server to four clients. In this work environment, suppose that the inventory, data entry, and executive clients create documents with graphics from the graphics file server. They connect to the graphics file server by a network drive to its host name. 18 System i: Networking Domain Name System

25 Figure 4. DNS and DHCP on the same System i platform Previous versions of DHCP and DNS were independent of each other. If DHCP assigned a new IP address to a client, the DNS records had to be manually updated by the administrator. In this example, if the graphics file server's IP address changes because it is assigned by DHCP, then its dependent clients will be unable to map a network drive to its host name because the DNS records will contain the file server's previous IP address. With the IBM i DNS server based on BIND 9, you can configure your DNS zone to accept dynamic updates to DNS records in conjunction with intermittent address changes through DHCP. For example, when the graphics file server renews its lease and is assigned an IP address of by the DHCP server, the associated DNS records are updated dynamically. This allows the other clients to query the DNS server for the graphics file server by their host names without interruption. To configure a DNS zone to accept dynamic updates, complete the following tasks: v Identify the dynamic zone You cannot manually update a dynamic zone while the server is running. Doing so might cause interference with incoming dynamic updates. Manual updates can be made when the server is stopped, but you will lose any dynamic updates sent while the server is down. For this reason, you might want to configure a separate dynamic zone to minimize the need for manual updates. See Determining domain structure for more information about configuring your zones to use the dynamic update function. Domain Name System 19

26 v Configure the allow-update option Any zone with the allow-update option configured is considered a dynamic zone. The allow-update option is set on a per-zone basis. To accept dynamic updates, the allow-update option must be enabled for this zone. For this example, the mycompany.com zone has allow-update data, but other zones defined on the server can be configured to be static or dynamic. v Configure DHCP to send dynamic updates You must authorize your DHCP server to update the DNS records for the IP addresses it has distributed. v Configure secondary server update preferences To keep secondary servers current, you can configure DNS to use the NOTIFY function to send a message to secondary servers for the mycompany.com zone when zone data changes. You should also configure incremental zone transfers (IXFR), which enables IXFR-enabled secondary servers to track and load only the updated zone data, instead of the entire zone. If you run DNS and DHCP on different servers, there are some additional configuration requirements for the DHCP server. Related concepts: Dynamic updates on page 5 IBM i Domain Name System (DNS) that is based on BIND 9 supports dynamic updates. Outside sources, such as Dynamic Host Configuration Protocol (DHCP), can send updates to the DNS server. In addition, you can also use DNS client tools, such as Dynamic Update Utility (NSUPDATE), to perform dynamic updates. Related tasks: Configuring the DHCP to send dynamic updates to DNS Related reference: Example: DNS and DHCP on different System i platforms Example: Splitting DNS over firewall by setting up two DNS servers on the same System i This example depicts a Domain Name System (DNS) server that operates over a firewall to protect internal data from the Internet, while allowing internal users to access data on the Internet. This configuration accomplishes this protection by setting up two DNS servers on the same System i platform. The following figure depicts a simple subnet network that uses a firewall for security. Suppose that the company has an internal network with reserved IP space and an external section of a network that is available to the public. The company wants its internal clients to be able to resolve external host names and to exchange mail with people on the outside. The company also wants its internal resolvers to have access to certain internal-only zones that are not available at all outside of the internal network. However, they do not want any outside resolvers to be able to access the internal network. With IBM i DNS based on BIND 9, you can use two ways to accomplish this. The first way is that the company sets up two DNS server instances on the same System i platform, one for the intranet and another for everything in its public domain, which is described in this example. Another way is to use the view function that is provided in BIND 9, which is described in the example about splitting DNS over firewall by using a view. 20 System i: Networking Domain Name System

27 Figure 5. Splitting DNS over a firewall by setting up two DNS servers on the same System i The external server, DNSB, is configured with the primary zone mycompany.com. This zone data includes only the resource records that are intended to be part of the public domain. The internal server, DNSA, is configured with the primary zone mycompany.com, but the zone data defined on DNSA contains intranet resource records. The forwarder option is defined as This forces DNSA to forward queries it cannot resolve to the DNSB server. If you are concerned about the integrity of your firewall or other security threats, you have the option of using the listen-on option to help protect internal data. To do this, you can configure the internal server to only allow queries to the internal mycompany.com zone from internal hosts. In order for all this to Domain Name System 21

28 work correctly, internal clients need to be configured to query only the DNSA server. You need to consider the following configuration settings to split DNS: v Listen-on In other DNS examples, only one DNS server is on a System i platform. It is set to listen on all interface IP addresses. Whenever you have multiple DNS servers on a System i platform, you must define the interface IP addresses that each one listens on. Two DNS servers cannot listen on the same address. In this case, assume that all queries that come in from the firewall are sent in on These queries should be sent to the external server. Therefore, DNSB is configured to listen on The internal server, DNSA, is configured to accept queries from anything on the 10.1.x.x interface IP addresses except To effectively exclude this address, the address match list must have the excluded address listed before the included address prefix. v Address match list order The first element in the address match list that a given address matches is used. For example, to allow all addresses on the 10.1.x.x network except , the ACL elements must be in the order (! ; 10.1/16). In this case, the address is compared to the first element and is immediately denied. If the elements are reversed (10.1/16;! ), the IP address is allowed access because the server compares it to the first element that matches, and allows it without checking the rest of the rules. Related reference: BIND 9 features on page 7 BIND 9 is similar to BIND 8; however, it provides several features to enhance performance of your Domain Name System (DNS) server, such as views. Example: Splitting DNS over firewall by using view This example depicts a Domain Name System (DNS) server that operates over a firewall to protect internal data from the Internet, while allowing internal users to access data on the Internet by using the view feature that BIND 9 provides. Example: Splitting DNS over firewall by using view This example depicts a Domain Name System (DNS) server that operates over a firewall to protect internal data from the Internet, while allowing internal users to access data on the Internet by using the view feature that BIND 9 provides. The following figure depicts a simple subnet network that uses a firewall for security. Suppose that the company has an internal network with reserved IP space and an external section of a network that is available to the public. The company wants its internal clients to be able to resolve external host names and to exchange mail with people outside the network. The company also wants its internal resolvers to have access to certain internal-only zones that are not available outside of the internal network. However, the company does not want any outside resolvers to be able to access the internal network. With IBM i DNS based on BIND 9, you can use two ways to accomplish this. The way described in this example is that you can configure the DNS server with two different views to listen on various queries, one for the intranet and another for everything in its public domain. Another way is to set up two DNS server instances on the same System i platform, which is described in the example about splitting DNS over a firewall by using two DNS servers. 22 System i: Networking Domain Name System

29 Figure 6. Splitting DNS over a firewall by using view The DNS server, DNSC, defines two views, called external and internal. The external view is configured with a primary zone mycompany.com that includes only the resource records that are intended to be part of the public domain, while the internal view is configured with a primary zone mycompany.com that contains intranet resource records. If you are concerned about the integrity of your firewall or other security threats, you have the option of using the match-clients substatement to help protect internal data. To do this, you can configure the internal view to only allow queries to the internal mycompany.com zone from internal hosts. You need to consider the following configuration settings to set up split DNS: v Match-clients Domain Name System 23